Professional Documents
Culture Documents
and GSM mobile RF 124), and is designed for wide area cellular opera-
tion with maximum output powers of 1 W to 8 W
allowed for mobile applications.
Out-of-channel measurements deter- maintained sufficiently with a mini- measurement is used to ensure that the
mine how much interference the user mum of power. This allows overall sys- modulation process does not cause
causes other GSM users. These mea- tem interference to be kept to a mini- excessive spread. This would cause
surements include spectrum due to mum, and in the case of an MS, battery interference to adjacent channel users.
modulation and wideband noise, spec- life is maximized. To perform these measurements,
trum due to switching, and Tx and Rx Power measurements outside of analyzers are tuned to spot frequen-
band spurious. specifications usually indicate a fault in cies and time gated across part of the
Out-of-band measurements deter- the power amplifier circuitry, the cali- modulated burst. Using this mode, the
mine how much interference the user bration tables or the power supply. power is measured. The analyzer is
causes other non GSM users of the The mean output power is mea- then retuned to the next frequency or
radio spectrum, such as the military, sured during the useful part of the another offset of interest. This process
police, and aviation. All other spurious GSM burst. When performing this continues until all offsets are mea-
(such as harmonics and wideband) are measurement, the GSM test equip- sured and verified against permissible
included here. ment derives the correct timing refer- limits. The result of these measure-
ence by demodulating incoming sig- ments is a set of frequency versus
Phase error and frequency error nals and gating over the useful part of power points that define the spectrum
Phase error is one of the parameters the GSM burst. of the signals. However, spectral com-
used in GSM to characterize modula- The transmitters of the GSM sys- ponents that result from the effect of
tion accuracy. Poor phase error usually tems must ramp up and down within bursting do not appear because the
indicates problems with I/Q baseband the time division multiple access ramps are gated out.
generators, filters, modulators, or (TDMA) structure to prevent adjacent The test limits for these measure-
amplifiers in the transmitter circuitry. timeslot interference. If transmitters ments are expressed in dBc (power
Frequency error measurements indi- turn on too slowly, data at the begin- below carrier). It follows that the first
cate poor synthesizer/phase lock loop ning of the burst might be lost, degrad- step of the measurement is to take a
performance (such as if synthesizers ing link quality. If the transmitters reading of the center frequency to
may not settle quickly enough as they turn off too slowly, the user of the next which the transmitter is tuned.
shift frequencies between transmis- time slot in the TDMA frame will expe-
sions). In GSM systems, poor frequency rience interference. Spectrum due to switching
error can cause target receivers to fail Therefore, transmitted RF carrier GSM transmitters ramp RF power
to gain lock to transmitted signals. Also power versus time measurements are rapidly. The transmitted RF carrier
the transmitter could cause interfer- performed to assess the envelope of car- power versus time measurements
ence with other users. rier power in the time domain against a described earlier ensure that this
To measure phase and frequency prescribed mask. The measurements process happens at the correct times
error, test sets can be used to sample also check that the transmitters’ turn and is fast enough. However, if RF
transmitted output of the devices under offs are complete. If a transmitter fails power is ramped too quickly, undesir-
test to capture the actual phase trajecto- this measurement, it usually indicates able spectral components exist in the
ry. This is then demodulated and the a problem with the units PA or power transmission. This measurement also
ideal phase trajectory is derived mathe- control loop. ensures that these components stay
matically. Subtracting one from the below the acceptable level.
other gives error signals. The mean gra- Adjacent channel power To perform spectrum due to switch-
dient of these signals gives frequency As part of the out-of-channel mea- ing measurements, the analyzers are
error. The variation of this signal is the surements, the adjacent channel tuned to and measure multiple offset
phase error and is expressed in terms of power (ACP) is defined by two mea- frequencies in zero span mode with no
root mean square (rms) and peak. Figure surements: spectrum due to modula- time gating.
2 demonstrates this test procedure. tion and wideband noise, and spec-
Figure 3 shows a measurement of trum due to switching. These two mea- Spurious measurements
the phase error on one transmitted surements are usually grouped togeth- The out-of-channel measurements
burst and how it relates to the limits er and referred to as output RF spec- are necessary to ensure GSM transmit-
set by the GSM standard. trum (ORFS). ters do not place energy into the incor-
The modulation process in a trans- rect parts of the spectrum. This would
Mean transmitted output power mitter causes continuous wave carriers cause interference to other users of the
GSM systems use dynamic power to spread spectrally. The spectrum due spectrum. These anomalies are referred
control to ensure that each link is to modulation and wideband noise to as spurious transmissions.
The spurious transmissions are mea- • Frame erasure rate (FER) — The are three key tests that define the
sured by connecting test sets directly to percentage of erased frames com- receivers blocking performance: spuri-
the antenna connectors of the MS. Due pared to the total number of frames ous immunity, intermodulation immu-
to the antennas direct connection to the sent during an observation period. nity, and adjacent channel selectivity.
test sets, these measurements are • Residual bit error rate (RBER) — Spurious immunity is the ability of
referred to as the conducted spurious When frames are erased, the BER the receivers to prevent single, out-of-
measurements. Measurements of this of the remaining frames is mea- channel interference signals from caus-
parameter include Tx/Rx band spuri- sured. The RBER parameter ing undesired in-channel responses at
ous, cross band spurious, and out-of- defines this measurement. the output of the receivers. Spurious
band spurious. BER is a ratio of bits received erro- may be generated within the receivers
The spurious measurements can be neously versus total number of bits from power supply harmonics, system
categorized as Tx or Rx depending on received. It is measured as follows: The clock harmonics or LO spurious.
the band they inhabit. The Tx band test systems output signals carrying Intermodulation immunity is a
spurious measurements relate to spuri- known bit patterns (usually pseudo measure of the receivers’ performance
ous that fall within the 925 MHz to 960 random bit sequences or PRBS). PRBS in the presence of distortion products
MHz GSM Tx band. The Rx band spu- signals are usually labeled PNx, where due to intermodulation products.
rious measurements, however, are x is the number of bits being permutat- These intermodulation products are
measures of how much energy the ed in the sequence (such as PN9 = 29 - 1 generated when more than one tone is
transmitters put in the Rx band (880 or 511 bits). present at the input of the receivers.
MHz to 915 MHz). This test ensures During the measurements, the The tones non-linearly mix to form
that Tx spurious don’t “jam” or desensi- receivers under test attempt to demodu- third-order intermodulation products.
tize adjacent receivers. The specifica- late and decode these patterns. By The products of concern lie within the
tions of the measurements are based on return paths (using a method known as receivers’ passband.
1 m average distance between mobiles. loop back), the receivers send the resul- Adjacent channel selectivity is a
For the purpose of attenuating the Tx tant bits back to the test systems for measure of the receivers’ ability to
band signals during these measurements, comparison. The test systems then calcu- process the desired modulated signals
the test setups usually include Rx band late the required metrics. GSM handsets in the presence of strong signals in the
pass filters in front of the analyzer inputs. are tested using this loop-back method. adjacent channels. Alternate channel
In some countries GSM900 and Most receivers are required to main- selectivity is a similar test in which the
DCS1800 systems co-exist. For this tain a specified BER in the presence of interfering signals are two RF channels
reason the ETSI 3GPP standards interfering signals within the channel. away from the receivers’ passband.
require specific cross-band performance For GSM, the performance is measured
capability. This is to ensure the GSM as follows: Digitally modulated signals GSM mobile RF transceiver
transmitters place the minimum ener- power levels are set 20 dBs above derivation
gy required into both the DCS1800 and receiver sensitivity at the center of the Receiver sensitivity is related to
GSM9000 bands. receivers’ passband. These signals are receiver noise figure according to:
The out-of-band spurious is a series combined with GMSK modulated inter-
of spectrum analyzer measurements ferers. Combined signals are then
over a large frequency range — from injected into the antenna ports of the
100 kHz through 12.75 GHz. The 3GPP receivers. The power levels of the Where the receivers’ bandwidth (180
standards were written to include GMSK interfering signals are then set kHz for GSM), is the baseband signal-
wideband spurious limits to which an to nominal levels at which the receiver to-noise-ratio and the RF and broad-
MS must conform. BERs must not exceed the receiver sen- band implementation gain.
sitivity specifications. The difference in The GSM standard specifies a mini-
Receivers power levels between the two signals is mum -102 dBm sensitivity requirement.
Sensitivity is the fundamental mea- the interference ratio. Given a worse-case baseband ratio of 9
sure of receiver performance. It speci- Receiver blocking constitutes one of dB and a 2 dB implementation margin,
fies the minimum signal level for a the out-of-channel receiver tests. we calculate the noise figure as:
specified percentage of errors in the Blocking tests verify correct receiver
demodulated information. The reported operation in the presence of out-of-
value for all receiver measurements is channel signals and monitor the
bit error rate (BER) or one of the fol- receivers’ susceptibility to internally Given this worse-case NF, the
lowing a variations: generated spurious responses. There receivers’ designers can then investi-
gate various front end gain and NF par- to achieve the same system NF as the
tition options according to the equation: dual LNA option.
The main disadvantage of the dual
LNA option typically is increased cost
and (potentially) the extra supply cur-
rent required due to addition of a sec-
ond LNA.
Where Fi is the noise factor of the
i’th block in the partition (i = 1, 2, 3 ...). Rx blocking analysis
Although the second equation shows Table 1 shows the blocking signal
that the higher the gain of the first levels with which a GSM mobile is
active stage, the lower the NF of the expected to perform without drop-
system would be, the receivers’ design- ping a call. GSM receiver designers
ers need to ensure that the first active specify the receive strips’ compres-
stage does not compress the subsequent sion points based on the above listed
stages. This would degrade receiver lin- in-band blocking signal level specifi-
earity. This shows that system sensitiv- cations, and uses the out-of-band
ity is a compromise between receiver blocking signal levels to define the
NF (dominated by choice of front end filter rejection specifications to avoid
components) and receiver linearity. signal path compression.
The receivers’ front end low noise For example, in-band blocking at 3
amplifier (LNA) options typically inves- MHz offset (such as -23 dBm) sets the
tigated are shown in figures 4 and 5. compression point required for the
The main benefit of the dual LNA front end. Assuming a 1 dB loss switch
option — when compared with the and a 2.5 dB loss filter prior to a LNA
first option — is that the individual stage in the receive strip, it places a
LNA noise figure and gain require- total 3.5 dB loss prior to the LNA stage.
ments are significantly relaxed. With This means the LNA compression point
the single LNA option, the front end must, in the worst case, be -26.5 dBm
LNA would need to be tightly specified (such as -23 dBm to 3.5 dBm).
where
Conclusion
This article is meant to only be an
overview of the key measurements
required for testing GSM transceivers,
and GSM mobile performance deriva-
tion. It is intended to help an RF
designer with no GSM RF system
knowledge get a better appreciation of
GSM module specifications and how
they can affect system performance.
RF Design www.rfdesign.com 21